
Work Here?
Fal.ai helps businesses improve data analytics using NLP and ML, focusing on sentiment analysis and anomaly detection within dbt data models. It analyzes text from customer reviews, support tickets, and surveys to label sentiment as positive, negative, or neutral, and flags unusual patterns in data transformations. The platform integrates with existing data infrastructure using dbt models and is offered via tiered subscriptions that include basic sentiment analysis, advanced anomaly detection, and premium support. Its goal is to help data-driven organizations make informed decisions, improve customer satisfaction, and get continuous analytics updates.
Industries
Data & Analytics
Enterprise Software
AI & Machine Learning
Company Size
51-200
Company Stage
Late Stage VC
Total Funding
$943.9M
Headquarters
Seattle, Washington
Founded
2021
See people who can refer or advise you
Help us improve and share your feedback! Did you find this helpful?
Total Funding
$943.9M
Above
Industry Average
Funded Over
8 Rounds
Health Insurance
Dental Insurance
Vision Insurance
Company Equity
Relocation Assistance
Fal, a generative media platform, has launched H3 Max, a new video generation model that ranks first on independent benchmarks from Artificial Analysis and Design Arena. The model generates five-second videos in approximately three seconds, achieving faster-than-real-time generation. Built on the open-weights MiniMax H3 model, H3 Max was developed through post-training for improved prompt adherence and visual quality. Fal reports the system delivers roughly 35 times the throughput of the official MiniMax H3 endpoint and averages 15 times faster than comparable quality models. H3 Max is available through the fal API and Playground. During a promotional launch period ending 7 September, pricing starts at $0.04 per second at 768p resolution, rising to $0.08 per second thereafter. Fal developed the model by co-designing training and inference optimisation, treating them as a single problem rather than separate layers.
Lore issue #200: OpenAI's first AI chip beats Nvidia's GB300 on efficiency. PLUS: Nvidia nears a $13B Hugging Face deal, Sam Altman says AGI could arrive this year, and Hugging Face unveils a $399 open-source robot Aug 28, 2026 Good morning, welcome to this week's Lore Brief, your 3-minute brief of the most important moves in AI and tech. This issue is brought to you by Factory, the fastest way to ship software with autonomous engineering agents. * OpenAI's Jalapeño chip beats Nvidia GB300 on efficiency | First published results show the custom inference chip delivering 1.5 to 1.9 times more work per watt than Nvidia's GB300 while cutting end-to-end latency by as much as 3.6 times. The 700-watt part goes up against a 1,400-watt flagship and still returns answers faster on open models like DeepSeek R1 and Kimi K2.5. Deployment inside OpenAI's own infrastructure is planned by year-end. Read more here | * fal's H3 Max claims the top spot in video generation | The post-trained model ranked first for quality prompt understanding and aesthetics against a dozen leading systems. It can produce a 5-second 720p clip in under three seconds and is 50 percent off this week. fal built it on MiniMax's open H3 base then tuned it for speed without giving up looks. Read more here | * Nvidia closes in on a Hugging Face acquisition | Talks point to a deal around $13 billion that would give the chipmaker a major foothold in open-source AI. Hugging Face last raised at $4.5 billion in 2023 and now does roughly $150 million in annual revenue. The move would help Nvidia stay central as closed labs build their own chips. Read more here Lip Sync tutorial for AI videos: The same style ad like the one of the new Mac mini but made with AI: * OpenAI leaders say AGI is getting close | Sam Altman told TIME the company could have an internal system that qualifies as AGI by the end of 2026. Mark Chen put the lab at about 80 percent of the way there. Astra already acts as an automated research intern that can run week-long experiments inside OpenAI's own codebase. Read more here | * A practical field guide to living with Grok Bot | Two weeks after launch Matt Van Horn published every hack he has found from giving a bot its own inbox to letting it place phone calls in Portuguese. The write-up covers planning layers named roles cookie sync and the habit of drafting before anything gets sent. The honest caveat is that unsupervised crews can multiply their own errors fast. Read more here | * TIME drops its 2026 list of AI's most influential people | The annual TIME100 AI ranking is out again with the usual mix of lab chiefs and public figures. Online reaction quickly turned to the oddities including Paris Hilton while Jensen Huang is missing from the list. Readers treated that combination as the punchline. Read more here | * Claude memory now follows you from chat into Cowork | One shared memory means a task in Cowork can start from what you already discussed in chat including projects preferences and past clients. You can read edit or delete every saved topic in Settings. Sensitive subjects stay out unless you turn them on. Read more here | * Hugging Face unveils a $399 open-source robot | Microduck can walk pick things up get back up after falling and even roller-skate. Users teach it new tricks with reinforcement learning instead of waiting for a closed lab to ship a firmware update. Clem Delangue frames it as affordable hardware for physical AI and world models. Read more here | * Skild AI's S1 learns 10-minute robot tasks from one video | Show the foundation model a single demonstration and it can complete jobs it never saw in training from making coffee to flipping pancakes. No fine-tuning is required and it can recover from mistakes the human in the video never made. The company says matching that accuracy with older VLAs would take 50 to 100 hours of extra data. Read more here | * Google ships Gemini 3.5 Transcribe for cleaner live speech | The new model turns messy audio into formatted text while cutting filler words and handling self-corrections like "Tuesday no Wednesday." It recognizes more than 85 languages and drops word error rates well below earlier Chirp systems. Developers can use it now through the Gemini API and it is already landing in the macOS Gemini app. Read more here That's it for this week's Lore Brief. See you next week!
MiniMax H3 Max: free fast AI video from fal. Fal has released MiniMax H3 Max, a post-trained version of MiniMax H3 tuned for speed: 5-second 768p clips with synced audio in under three seconds. Fal has released MiniMax H3 Max, a post-trained version of MiniMax's open-weights H3 video model tuned for speed. It renders a 5-second, 768p clip with synchronized audio in under three seconds, and it landed at the top of the image-to-video arena the week it shipped. Try it: five free clips a day. Sign in to fal and you get five free H3 Max generations every 24 hours, each up to 15 seconds at 768p with synced audio. Open the text-to-video endpoint, write a prompt, and pick an aspect ratio from 21:9 down to 9:16 for vertical shorts. Because a 5-second clip comes back in roughly the time it takes to read this sentence, you can iterate on shot ideas in a single sitting instead of queuing renders and walking away. Why it matters for creators. H3 Max is not a new base model. Fal took the open weights of MiniMax H3 and retrained them for stronger prompt adherence and better aesthetics, then served the result on its own inference stack. The trade is resolution for speed: H3 Max caps at 768p where the base model reaches 2K, but it generates fast enough to feel interactive. For creators who tested audio-synced generation in its ComfyUI H3 sound-sync coverage, this is the same model family with a free, hosted, near-real-time front end. Key details. Base model: Post-trained from MiniMax H3 open weights by fal, developed by MiniMax. Resolution: 480p or 768p (default 1344x768 at 24 FPS), tuned for speed over maximum resolution. Speed: A 5-second 768p clip renders in under three seconds; 15-second clips take about 15 seconds. Free tier: Five 15-second generations every 24 hours with synchronized audio. Pricing: $0.06 per second at 768p, with a 50 percent introductory discount for the first 14 days. Rankings: First for image-to-video on Design Arena and first with audio on the Artificial Analysis video leaderboard. What to do next. Run the same prompt through H3 Max and your current video tool and compare prompt adherence and motion. If the 768p ceiling works for social shorts and rough cuts, the free daily quota makes it a low-cost way to storyboard before committing paid credits on a higher-resolution model for the final render.
fal announced fal Agent, a creative AI partner that orchestrates production-ready workflows across leading image, video, and 3D generative models. The tool maintains context and consistency throughout projects, allowing creatives to move between different models without rebuilding prompts or losing creative decisions. fal Agent automatically selects appropriate models for each task whilst preserving characters, objects, and visual styles across generations. Projects retain memory and references, enabling teams to resume work weeks later with full context intact. The platform integrates with fal's API and CLI, making it accessible for developers building AI-powered creative tools. It runs on fal's inference infrastructure, trusted by over 2.5 million developers. fal Agent is now available in early access at fal.ai/agent, with add-on credits usable across the entire fal platform.
fal adds LoRA training for MiniMax H3, starting with an open-source Realism People LoRA. Custom LoRA training for MiniMax H3 is now live on fal, and the company demonstrated it with Realism People, an open-source LoRA it says pushes the model toward photorealistic humans. * The trainer targets MiniMax H3. It fine-tunes the open-weight model that generates video, image, audio, and text. * Realism People ships open source. fal posted the LoRA weights to Hugging Face, tuned for skin, eyes, and motion. * The trainer covers several input modes. fal points to entry points for text-to-video, image-to-video, first-last-frame, and reference-to-video. * More LoRAs are coming. fal says additional LoRAs are on the way. Realism People targets photorealistic humans. fal trained Realism People to show what the new trainer produces, describing it as a LoRA that pushes H3 toward raw, photorealistic humans with a focus on skin, eyes, and motion. fal posted the weights as an open-source release on Hugging Face, so you can run the LoRA directly or study how it was built before training your own. Because it is an open release rather than a hosted-only feature, the LoRA can be pulled into other workflows built around H3 rather than staying locked to fal's interface. VP Land has also compared reference images versus LoRAs for holding a consistent look across shots. Training on an open-weight video model. A LoRA lets you steer a base model toward a specific look, subject, or style without retraining the whole thing, which is how fal can layer training on top of H3. If the term is new to you, VP Land broke down what a LoRA is for non-technical readers. fal lists trainer entry points spanning text-to-video, image-to-video, first-last-frame, and reference-to-video, so a custom LoRA can attach to more than one generation path. Teams wanting a consistent character or aesthetic across shots can tune H3 on their own material and reuse the result instead of prompting it blind each time. A trainer and reference LoRA make H3 customizable. The MiniMax H3 launch gave creators an open-weight model that unifies multiple generation types. Adding a trainer and a reference LoRA turns that base model into something you can customize, and the open-source Realism People release gives a concrete starting point rather than a closed demo. With fal signaling more LoRAs to follow, the near-term value is a growing library of tunable looks for anyone already generating with H3.
Find jobs on Simplify and start your career today
Industries
Data & Analytics
Enterprise Software
AI & Machine Learning
Company Size
51-200
Company Stage
Late Stage VC
Total Funding
$943.9M
Headquarters
Seattle, Washington
Founded
2021
Find jobs on Simplify and start your career today